Palliative Medicine 2014, Vol. 28(4) 335 –341© The Author(s) 2014Reprints and permissions:
Increasing anticholinergic burden and delirium in palliative care inpatients
Kristin M Zimmerman 1,2, Marci Salow 2,3, L Michal Skarf 3,4, Tia Kostas 3,4,5, Allison Paquin 2,3, Mark J Simone 4,5 and James Rudolph 3,4
Abstract
除腋下臭味
Background: Delirium may complicate the hospital cour and adverly impact remaining quality of life for palliative care inpatients. Medications with anticholinergic properties have been linked to delirium within elderly populations via rum anticholinergic assays.
Aim: The aim of this study is to determine whether increasing anticholinergic burden, as measured using a clinical asssment tool, is associated with an increa in delirium among palliative care inpatients.Design: This study was completed as a retrospective, ca-control study.
Setting/participants: Veterans admitted to the Veterans Affairs Boston Healthcare System and consulte
d to the palliative care rvice were considered for inclusion. Increa in anticholinergic burden from admission through hospital day 14 was assd using the Anticholinergic Risk Scale. Prence of delirium was determined by u of a validated chart review instrument.
Results: A total of 217 patients were analyzed, with a mean age of 72.9 (±12.8) years. The overall delirium rate was 31% (n = 67). Patients with an increa in Anticholinergic Risk Scale (n = 72 (33%)) were 40% more likely to experience delirium (odds ratio = 1.44, 95% confidence interval = 1.07–1.94) compared to tho without increa (n = 145 (67%)). After adjustment for age, brain metastasis, intensive care unit admission, illness verity, opiate u, and admission Anticholinergic Risk Scale using multivariable modeling, delirium risk remained significantly higher in patients with an Anticholinergic Risk Scale increa compared to tho without increa (adjusted odds ratio = 1.43, 95% confidence interval = 1.04–1.94).
Conclusion: An increa in Anticholinergic Risk Scale from admission was associated with delirium in palliative care inpatients. While additional study is needed, anticholinergic burden should be incread cautiously in palliative inpatients, and tho with increas should be cloly followed for delirium.
Keywords
Delirium, anticholinergic, palliative care, aged, psychopharmacology
Original Article
1 D epartment of Pharmacy Practice, Massachutts College of Pharmacy & Health Sciences University, Boston, MA, USA
2 D epartment of Pharmacy, Veterans Affairs (VA) Boston Healthcare System, Boston, MA, USA
3 D ivision of Geriatrics and Palliative Care, Veterans Affairs (VA) Boston Healthcare System, Boston, MA, USA
4Division of Aging, Harvard Medical School, Boston, MA, USA 5Division of Aging, Brigham and Women’s Hospital, Boston, MA, USA
Corresponding author:
Kristin M Zimmerman, Massachutts College of Pharmacy & Health Sciences University, 179 Longwood Avenue, Boston, MA 02115, USA. Email: kristin.zimmerman@mcphs.edu
336 Palliative Medicine28(4)
Introduction
Delirium is an acute fluctuation in cognition occurring with either a disturbance of consciousness or impairment in perception. Delirium occurs in 10%–30% of medically ill, hospital inpatients and is associated with increas in morbidity and mortality.1 Delirious patients are at incread risk of medical complications, prolonged hospitalizations, functional decline, and institutionalization.1
Palliative care patients are at high risk of delirium, where as many as 83% of terminally ill patients will develop delir-ium near death.1 In this population, delirium has been shown to contribute to patient, family, and caregiver stress.2,3 The ability to remain oriented to surroundings, with prerve
d cognitive function and ability to communicate, is important for patients and caregivers at the end of life.4–6 Disturbance in consciousness, changes in cognition, and impairments in perception associated with delirium may interfere with the experience of “dying well.”4–6
The caus of delirium in palliative care are typically multifactorial, including metabolic dysfunction, brain metastas, infections, and medications.7,8 Medications with anticholinergic properties are particularly concerning as rum anticholinergic activity has been correlated to delirium incidence and verity in elderly patient popula-tions, with reductions linked to symptom improvement.9–12 Anticholinergics have also been linked to cumulative risk of cognitive impairment and mortality in such popula-tions.13 Such links between rum anticholinergic levels and delirium have not been demonstrated within the pallia-tive care population. Additionally, u of rum anticholin-ergic assays is impractical in the cour of clinical care, and its associations with true medication load questioned.14 While clinical tools to quantify anticholinergic burden exist, they have not been utilized to predict or characterize delirium occurrence.
Drugs with anticholinergic properties are often ud in palliative care for end-of-life symptoms and management of comorbid conditions. Unfortunately, desired physio-logic effects, such as anti-cretory, anti-emetic, and anti-diarrheal, may be accompanied by deleterious cognitive effects. Exa
mples of anticholinergic medications include atropine, scopolamine, ipratropium, ranitidine, promethaz-ine, and loperamide. In their study of palliative care outpa-tients in Australia, Agar et al.15 found that increas in anticholinergic burden are associated with decreas in functional measures and quality of life as patients approach death. Anticholinergic side effects found in their popula-tion included incread dry mouth and incread difficulty concentrating. However, there was no increa in confu-sion or hallucinations.15 Caring for palliative care patients at the end of life necessitates a judicious balance between the benefits of symptom management afforded by the u of anticholinergic medications and the significant risks of delirium from the medications. Since the clinical deci-sion is often difficult, additional information regarding the overall risk of increasing anticholinergic burden would be helpful. In addition, the ability to predict the time cour for the ont of delirium is also helpful in this population, particularly when patients may have a prognosis of days to weeks.
Validated tools to clinically quantify this anticholiner-gic burden have been developed. The Anticholinergic Risk Scale (ARS) is a tool designed to estimate the extent to which an individual may be at risk of anticholinergic adver effects that can lead to cognitive dysfunction and delirium (Figure 1). The scale ranks medications accord-ing to their anticholinergic potential on a 3-point scal
e (0 (no or low risk) to 3 (high anticholinergic potential)).16 The total risk is equal to the sum of the points for all medi-cations. Higher ARS scores were significantly associated with incread risk of anticholinergic adver effects in geriatric outpatient populations.16
I t remains unknown how anticholinergic medications may contribute to delirium in the palliative care population. While prence or abnce of anticholinergic burden, as measured by rum anticholinergic assays, has been associ-ated with delirium in geriatric populations, correlations with clinical measures of anticholinergic burden have not been made within the palliative care population. To our knowl-edge, only two investigations have utilized clinical tools for the asssment of delirium within older inpatients.17,18 Han et al.17 utilized the Summers’ Drug Risk Scale to associate delirium verity with anticholinergic medications in
Zimmerman et al. 337
already delirious older medical inpatients. Lafortune et al.18 utilized the ARS to associate incread values with delirium in cognitively impaired older inpatients. Furthermore, the effect of an increa in anticholinergic burden on delirium occurrence, timing, and modification has not been studied. Given that anticholinergic medications may alleviate many end-of-life symptoms, it is critical to have a more in-depth understanding of anticholinergic medications’ relationship to delirium in palliative care patients to allow providers to weigh the risks and benefits of using the medications as death approaches. U of clinical tools such as the ARS to characterize this relationship can have significant clinical impact to reduce delirium by better encouraging u of alternative agents and delirium reduction measures at bur-den thresholds.
This study aimed to determine whether an increa in anticholinergic burden throughout admission, as measured using a clinical tool, the ARS, was associated with incread odds of delirium in palliative care patients. Secondarily, we aimed to characterize the temporal rela-tionship between increas in anticholinergic burden and delirium ont. To achieve this, we retrospectively reviewed palliative care inpatients and assd changes in ARS scores throughout hospitalization to asss for impact on odds of delirium.Materials and methods
Sample
This study retrospectively reviewed patients admitted to the Veterans Affairs (V A) Boston Healthcare System. Patients admitted between 1 July 2008 and 1 January 2010 and evaluated by the palliative care team during their admission were considered for inclusion. Patients were excluded from the study if the u of antipsychotics prior to admission was documented within the patient’s medica-tion list in the computerized patient record system (CPRS) or in the history and physical admission note. Patients with evidence of delirium or mental status changes documented in history and physical or nursing admission notes were also excluded. The protocol was approved by the V A Boston Healthcare System’s Institutional Review Board. Data collection
Data were collected from the time of hospital admission (hospital day 0) through index date (no later than hospital day 14) from the electronic medical record. Index date was defined as the date of outcome occurrence (development of delirium) if outcome occurred prior to day 14 of admis-sion. If no outcome occurred prior to day 14 of admission,
1 Point
2 Points
3 Points
Carbidopa-levodopa Entacapone Haloperidol Methocarbamol Metoclopramide Mirtazapine Paroxetine Pramipexole Quetiapine Ranitidine Risperidone Selegiline Trazodone Ziprasidone Amantadine
Baclofen
Cetirizine
Cimetidine
Clozapine
Cyclobenzaprine
Desipramine
不到黄昏梦未成打一字Loperamide
Loratadine
Nortriptyline
Olanzapine
Prochlorperazine Pudoephedrine
Tolterodine
Amitriptyline
Atropine products
Benztropine
Carisoprodol
Chlorpheniramine
Chlorpromazine
Cyproheptadine
Dicyclomine
Diphenhydramine
Fluphenazine
Hydroxyzine
Hyoscyamine
products
Imipramine
Meclizine
Oxybutynin
Perphenazine
心理学pptPromethazine
Thioridazine
Thiothixene
Tizanidine
Trifluoperazine
Figure 1.Anticholinergic Risk Scale. Total patient Anticholinergic Risk Scale score is defined as the sum of points for all medications.
338 Palliative Medicine 28(4)
the index date defaulted to day 14. If the patient died prior to day 14, the index date defaulted to the date of death. If multiple outcomes occurred, only the first outcome was labeled as the index date and included in analysis.
Exposure
Exposure was defined as the change in anticholinergic bur-den from admission to index date and was quantified using the ARS (e Figure 2). ARS scores were calculated manu-ally from data contained within the electronic medical record, utilizing the outpatient prescription record and history and physical examination notes for calculation of admission ARS and utilizing the computerized medication administra-tion record (MAR) for calculation of peak ARS. The peak ARS was defined as the highest total calculated value of administered anticholinergic medications at any time prior to the index date. If medications were administered on an as-needed basis, the MAR was queried to verify administration. Cas were defined as patients who had an increa in the ARS from admission to index date, whereas controls included tho who had no change or a decrea in ARS.
Outcome
Records were reviewed for delirium using a validated chart-bad instrument developed by Inouye et al.19 This chart-bad instrument prompts reviewers to arch for and locate key terms and descriptors for identification of delirium. A positive result is indicated by the prence of the key terms and descriptors in addition to evidence of acute ont or acute change in symptoms.19
Covariates
Additional data collected included basic demographics, such as age and gender. Consistent with previous delirium literature, data to complete an Acute Physiology and Chronic Health Evaluation–III (APACHE-III) score on all subjects were collected to account for potentially confounding vari-ables.20 Additional data collected included prence of brain metastasis, comorbid dementing illness, and evidence of metabolic dysfunction or infection.
精彩绝伦什么意思Statistical analysis
Bivariable analysis compared the characteristics of patients with and without an increa in ARS from time of admis-sion to index date. Bivariable associations ud a t -test for continuous variables and a chi-square test for dichotomous variables. For adjusted analys, logistic regression was ud to determine the odds for developing delirium. All variables with a p < 0.20 in bivariable analysis were entered into the model. Additionally, age and APACHE-III were included in the model, becau of their known asso-ciations with delirium.18 Odds ratios (ORs) for delirium are reported with 95% confidence intervals (CIs).
Additionally, we examined trends for delirium bad on the change in ARS according to categories (reduction, no change, increa of 1–2 points, and increa of ≥3 points). This trend was examined
with a chi-square for trend test. All statistical analys were conducted using STATA 11.0 (STATA, Inc., College Station, TX, USA).
Results
The 217 patients included in this study had a mean age of 72.9 (±12.8) years, and 97% were male. Of the initial pal-liative care consults, 50% were excluded due to prence of antipsychotic u or mental status change at the time of admission. The overall rate of delirium in this cohort was 31% (n = 67). Table 1 compares characteristics of patients with and without an increa in ARS. The exposure groups did not differ significantly with regard to gender, dementia diagnosis, APACHE-III scores, or opiate u. Tho with an increa in ARS were significantly more likely to be admitted to the intensive care unit (I CU) during their admission. Occurrence of delirium was significantly more common in patients with an increa in ARS (n = 72) com-pared to tho with no increa (n = 145): 42% versus 26%, p = 0.01.
Patients with an increa in ARS had a lower ARS score at admission of 0.8 ± 1.7 compared to 1.2 ± 1.7 in tho with no increa (p = 0.10). Patients with an increa in ARS had the highest peak mean ARS value throughout admission: 3.0 ± 2.1 compared to 0.8 ± 1.4 in tho with no increa in
ARS. In patients with an increa in ARS, peak occurred on hospital day 4.2 ± 3.2, and magnitude of change was 2.2 ± 1.2. Within this group, delirium occurred on hospital day 6.5 ± 3.4, 2.3 ± 3.2 days after the peak in ARS.
元旦作文300字As shown in Table 2, any increa in the ARS signifi-cantly incread the odds of delirium (OR = 1.4, 95% CI = 1.1–1.9) in bivariable analysis. After multivariable adjust-ment, an increa in ARS remained significantly associated with incread odds of developing delirium (adjusted OR = 1.4, 95% CI = 1.0–1.9). Table 3 illustrates the association of delirium with increasing magnitude of ARS changes. Overall, there was a trend toward incread delirium with magnitude of increa in ARS (chi-square = 5.8; p
= 0.02).
Figure 2. Study protocol. Peak Anticholinergic Risk Scale奥运主题歌
(ARS) score values were assd prior to index date and were compared with ARS values at admission.
Zimmerman et al. 339
Discussion
This retrospective study sought to determine whether an increa in anticholinergic burden, as measured by the ARS, was associated with an increa in delirium in pallia-tive care inpatients. The findings demonstrate that an increa in ARS was associated with statistically signifi-cant incread odds of delirium in this population. Compared to tho with no increa, tho with an increa in ARS after hospital admission were associated with a 40% increa in odds of delirium. While the findings demonstrate an association, our study design prevents causal attribution. Increas were equivalent in magnitude to approximately 2 points on the ARS—similar in value to the addition of medications such as loratadine or lopera-mide—and peaked approximately 2 days prior to delirium occurrence. I n patients experiencing delirium, the peak ARS values and the degree of ARS change were signifi-cantly higher than tho who did not experience delirium. By quantifying and characterizing the incread odds of delirium, the findings highlight the need for greater attention to, and judicious u of, anticholinergic medica-tions at the time of prescription. This study also showcas the u of a practical clinical tool, the ARS, for such asssment.
The association of delirium and anticholinergic load found in this study is consistent with prior work.9–12,17,18 As anticholinergic assays ud in prior studies may be diffi-cult to obtain and clinica
lly interpret, the utilization of a validated tool within this study may have practical impli-cations as noted above. The findings validate concern for deliriogenic potential of anticholinergic medications in
Table 1.Characteristics of exposure groups.
Characteristic No increa in ARS (n = 145)Increa in ARS (n = 72)p value Age, mean (±SD)74 (11)70 (15)0.04 APACHE-III, mean (±SD)45 (12)43 (13)0.37 Admission ARS, mean (±SD) 1.2 (1.7)0.8 (1.7)0.10 Male, % (n)97% (141)96% (69)0.58 Brain metastasis, % (n)5% (6)1% (1)0.15 Dementia, % (n)3% (6)3% (2)0.61 ICU admission, % (n)25% (37)42% (30)0.04 Opiates on admission, % (n)34% (49)29% (21)0.69 Change in opiates, % (n)49% (68)55% (38)0.40 SD: standard deviation; APACHE-III: Acute Physiology and Chronic Health Evaluation–III score at admission; admission ARS: Anticholinergic Risk Scale at admission; dementia: prence of International Classification of Dias (ICD-9) code consistent with dementia diagnosis; ICU admission: intensive care unit visit prior to index date; opiates on admission: u of opiates documented in medication lists or history and physical at admission; change in opiates: change in prescribed opiate regimen throughout admission.
Table 2.Adjusted ORs for delirium.
关于教育的文章
Bivariable associations with delirium Multivariable associations with delirium
OR95% CI Adjusted OR95% CI Age (per year) 1.010.99–1.03 1.010.99–1.04 Brain metastasis0.30.03–2.20.30.04–2.9 APACHE-III (per point) 1.010.99–1.04 1.00.98–1.04 ICU admission 1.50.8–2.8 1.40.7–2.6 Increa in ARS 1.4 1.1–1.9 1.4 1.0–1.9 OR: odds ratio; CI: confidence interval; brain metastasis: prence of brain metastasis documented in admission history and physical; APACHE-III: Acute Physiology and Chronic Health Evaluation–III score at admission; ICU admission: intensive care unit visit prior to index date; increa in ARS: change in Anticholinergic Risk Scale score from admission.
Table 3.Delirium with ARS increa.
ARS change
category (n)
Degree of ARS change Delirium, n (%)
Decrea (27)−1.9 (±0.7) 6 (22)
No change (118)0 (±0)31 (26)
Increa = 1–2
points (43)
1.3 (±0.5)17 (40)
Increa ≥ 3
points (29)
3.4 (±0.8)13 (45)
ARS: Anticholinergic Risk Scale.
Patients were divided into ARS change category depending on the
decrea, abnce of change, or increa by 1–2 or ≥3 points on the
ARS. Degree of ARS change reflects the mean change in ARS for each
patient group.
340 Palliative Medicine28(4)
this population and support a temporal, potentially predict-able and additive relationship between increasing anticho-linergic load and delirium.
There have been studies with findings contrary to the. However, the studies had important differences that may explain the disparate findings. One study by Campbell et al.21 of older, cognitively impaired, hospitalized adults did not find an association with anticholinergic medica-tions and delirium. This study, however, quantified anticholinergic burden utilizing an alternate scale, the Anticholinergic Cognitive Burden Scale. I n addition, Gaudreau et al.’s22 study of hospitalized oncology patients also failed to find an association between anticholinergic medications and delirium. The study by Gaudreau et al. did not u a validated tool for anticholinergic asssment. Neither study assd for the impact that changes in anticholinergic exposure may have on delirium. Finally, van Munster et al.23 found that predisposing factors for delirium co-occur with rum anticholinergic assays and may neutralize the associations between anticholinergics and delirium. However, as with many medications, the pharmacokinetic and pharmacodynamic changes, includ-ing inflammation, may result in enhanced anticholinergic impact at similar assay levels.23
The comprehensive asssment of exposure is a strength of this study. The CPRS allowed access to point-of-care MARs and nursing asssments. Utilization of this system allowed for the inclusion of important covari-ates, including confounding medications, laboratory and asssment values, and the accounting of administration and measurement timing. This system permitted the com-plete asssment of otherwi difficult-to-measure data.
Challenges to this work remain. The multifactorial nature of delirium, its under-recognition, and the inability to completely control for confounders may be viewed as limitations. For example, 55% of the control group devel-oped delirium despite no change, or even a decrea, in ARS. The APACHE-I I I score is ud to limit this con-founding within the delirium literature and was included in our analysis. The APACHE-I asssment for dia verity accounts for the delirium risk factors of age, ill-ness, and infection verity. Opioids were the only poten-tially confounding medications assd in this study. Limitations to delirium asssment are inherent in this study’s widely male population and retrospective, obr-vational nature. The ORs ud in this ca-control study can only approximate the relative risk for delirium with anticholinergic medication u. The nsitivity and speci-ficity of ca and control identification rely on that of the chart review tool, which is subject to interpretation bias; however, the tool designed by Inouye et al.19 was validated with a nsitivity of 74% and a spe
cificity of 83%. Finally, the effect of the exposure may be underestimated by the inability to account for delirium occurrence after the index date.
Delirium is common in palliative care patients, yet can be reverd up to 50% of the time.9,24 This study high-lights the association of delirium and increasing anticho-linergic burden in palliative care inpatients. The association of incread delirium odds with ARS increas of approximately 2 points underscores the importance of judicious anticholinergic prescribing in this vulnerable patient population. As prior studies have demonstrated that delirium is often under-recognized and detected late in the cour, the finding that peak ARS precedes clinical delirium by approximately 2 days may justify preemptive anticholinergic medication screening and asssment, and emphasize cautionary anticholinergic prescribing and minimization of other delirium risk factors in this popula-tion.1 In patients with symptoms warranting anticholiner-gic medication but with preexisting high ARS scores, delirium risk can be minimized by decreasing other deliri-ogenic medications, including opioids, and benzodiaz-epines, and by providing nonpharmacologic treatments such as hypnosis and behavioral therapies, which can be effective for the treatment of naua and insomnia, or by using anticholinergic medications with a lower anticho-linergic burden such as glycopyrrolate, haldol, or trazo-done, or in lower dos. I n addition, frequency of screening for delirium can be incread in p
atients with high ARS scores. Finally, using the ARS score to predict delirium allows clinicians to prepare families for the dif-ficulty this will cau.
This study builds a foundation for u of the ARS, in addition to the asssment of other important delirium risk factors, in clinical care decisions and screening measures to support the early screening, identification, and possible prevention of delirium. Future u may include systems-level calculations of ARS to target high-risk patients. Further prospective study is needed to confirm a causal relationship.
Conclusion
笔记本显卡坏了In this retrospective review, we found that an increa in anticholinergic risk scores after hospital admission was associated with a 40% increa in odds of delirium among palliative care inpatients. This obrved link between anticholinergic burden and delirium warrants future inves-tigation in a prospective study in order to further under-stand the true cau and effect relationship. I ncrea in ARS (18% vs 13%, p = 0.31) or occurrence of delirium (22% vs 13%, p = 0.09) was not significantly associated with an increa in death within 14 days. In an effort to minimize delirium, u of clinical tools such as the ARS may be ud to weigh risks and benefits associated with anticholinergic medi
cations in this vulnerable cohort. Future work linking ARS and delirium may prove uful in the prediction, prevention, and early intervention of delirium.